During the mid 1990s, national guidelines were established in accordance with World Health Organization recommendations for the diagnosis of uncomplicated malaria in Bangladesh. Based on simple clinical and epidemiologic criteria these guidelines were designed to be applied outside of tertiary care centers where microscopy was not feasible. We evaluated the positive predictive value (PPV) of these criteria using microscopic slide examinations as the gold standard in 684 subjects diagnosed and treated for malaria, sampling from eight subdistrict centers. The PPV for malaria was 32% with 19% for falciparum and 14% for Plasmodium vivax. Medical officers assigned to the study also gave their own clinical impression of whether cases could have been malaria. With the additional criteria of a medical officers' diagnosis, the PPV increased negligibly to 37% with 23% and 14% for falciparum and vivax, respectively. Since the PPV of diagnosis is low and cannot be improved on clinical grounds alone, we recommend the incorporation of laboratory diagnosis. This is especially important as we detect resistance to the first-line therapy chloroquine and require more expensive, potentially more toxic, regimens.